Claude Tools / Plugins / caveman
Audited plugin listing
caveman
Compresses Claude Code output using caveman-style speech — saves tokens in chat replies without touching code.
Listing signals
Install check
✓ Pass
Risk scan
✓ Clean
Implementation
Code-backed
Assessment
Verdict
Caveman is a tightly scoped, well-documented token-compression skill with clean hooks, zero runtime network calls, and benchmark numbers that are more carefully qualified than the headline suggests. The 65% token reduction is real for prose replies; the JetBrains-measured 8.5% applies to agentic coding runs, and the README explains both in the same paragraph — an unusual level of honesty for a plugin README. Hooks are purely local, install is standard, and the caveman-stats command gives users an actual receipt of session savings rather than a marketing promise.
Installation
Add to Claude Code
Source: github.com/JuliusBrussee/caveman
Evidence
Audit receipts
This listing is tied to a specific source commit and records both installation and risk checks.
Install check
✓ Pass
Has .claude-plugin/ with marketplace.json and plugin.json. SessionStart and UserPromptSubmit hooks via Node 18+.
Risk scan
✓ Clean
Composition
What was measured
Prose lines
4,234
Code lines
11,312
Prose / code
0.37
Similarity
Cluster note
Several language-specific ports exist (cavemanov for Russian/Kazakh, others). This is the English original and the best-maintained.
Related